Wednesday, April 21, 2010

DNA Computing

Computers, by definition, are machines which receive input, manipulate and store the input, and produce an output. They've quickly grown in the size and processing power. Computers are commonly known to consist of integrated circuits mainly constructed of silicon; however, a computer is never considered to be "alive." Technological advances however could use the building blocks of our genome in creating computer processors and data storage, and catapult processing speeds to incomprehensible levels not possible by today's standards.

DNA computing is an alternative to the way computers work today. While this technology is not readily available, or being mass produced, the theory behind it is quite old and the development is ongoing and catching more speed. Companies like IBM are attempting to use DNA to produce the next generation of processors.

Before discussing how DNA can be used in computers, it's important to first understand the basic structure of a DNA molecule. DNA is a double stranded helix where the two strands are linked by base pairs of amino acids commonly labeled, A, T, C, and G. A single double helix strand has millions of these connections which are limited as they only connect A to T and C to G. These amino acids would essentially take the place of the binary code of 1's and 0's used in computers today. The base pairs of amino acids are separated by.33 nanometers. To put the size into perspective, a DNA chip can be built in a 2-nanometer scale, when currently the top of the line chip is built with a 45-nanometer node. To put this another way, there are about 1 million gigabits of data per square inch of DNA. That figure explains why the idea is being developed by companies like IBM since this is 2000 times greater than our current data storage systems. Jennifer Cha, as biochemist at IBM indicated that "There is nothing else out there that we can do that with."

In 1994, Leonard Adleman, a professor at the University of Southern California, first introduced this theory in an article in the journal Science. Adleman discussed using the DNA molecule in computing, and demonstrated how it can be used to solve a seven point Hamiltonian path problem.

In short, a Hamiltonian Path is a traceable path that visits each vertex, or point, once with a beginning and ending point. While this may seem simple in theory, it is actually a complex problem to solve. To simplify this, if one was to try and plot the shortest route to tour the ten biggest cities in the United Kingdom only once, over 3.5 million routes would need to be analyzed. If this example were tested using a single processor each of the 3.5 million scenarios would need to be calculated one at a time, and the Hamiltonian Path would then be selected.

Rather than use ten vertexes Adleman used seven for his experiment. He encoded all the possible solutions, both correct and incorrect, in a large number of DNA. He labeled each of the seven cities in a four character combination of the base pairs of amino acid found in DNA. An example of one city indicator would be TCGG. By mixing all of the molecules in a test tube, he created all the DNA combinations, or answers, possible for the given conundrum. In theory this allowed for simultaneous processing in order to find the correct solution, as the DNA strands were developing not in succession of one another, but at the same time. Through a series of chemical reactions, Adleman was able to remove the incorrect answers and was able to leave only those strands representing the correct Hamiltonian path.

Taking this one step further, a team of United States scientists from universities across the nation, recently were able to engineer bacteria in order to solve a Hamiltonian Path Problem. Their research was reported in a July 2009 article in the Journal of Biological Engineering. Progressing Adleman's original research, their use of a biological computer allowed for the processing capacity to continue increasing through the process of cell division.

The scientists in this experiment used three vertexes to perform their analysis. Using the bacteria Escherichia coli, commonly referred to as E-coli, the vertexes were assembled using a combination of genes resulting in either a fluorescing green or red color. By randomly shuffling the DNA, the correct answer, or route, would cause the bacteria to glow both colors, giving it a yellow color. The results were verified by the scientists by ensuring the DNA sequence of the yellow bacteria was a result of genotypes representing a Hamiltonian path.

As can be seen with the bacterial computing experiment and Adleman's experiment, there are several flaws with DNA computing. A major drawback is the need for human intervention. Programming the inputs for the DNA computer is a complicated process. The scientists responsible for building the bacteria computer were required to first encode the DNA structure to reflect the 3 random vertexes. Adleman had to first create DNA strings to represent the 7 vertexes for his experiment. The analysis of the output as well requires human interpretation. Adleman was required to create a series of chemical reactions to abstract the Hamiltonian path strings, while the scientist in the biological computer experiment needed to check for yellow bacteria.

So what exactly could a DNA computer do? Obviously scientists are not simply trying to find the best way to solve a Hamiltonian path problem, nor is it planned that the home computer is to be replaced by DNA powered processors. The concept of DNA computing is a highly debated topic. The research and development done thus far could open our world up to a new class of computing devices. One possibility discussed is being able to apply tiny DNA computers inside of the body to help monitor and prevent diseases. The computer would analyze conditions, and make decisions based on their findings. Theoretically the tiny DNA computer would be able to release medicine or kill diseased cells. The new processors could also take the place of the current day supercomputers used for data crunching in large corporations, scientific labs, and government agencies. Processors faced with computing year's worth of data could cut the processing time into a fraction of what if currently takes.

DNA could also prove to be a much cheaper alternative to our current data storage technology. One gram of genetic material, which is the size of one cubic centimeter, could hold the equivalent of 1 Trillion compact discs. In a 2002 article of Business Week the estimated cost of a DNA sequence needed for computing was $30, compared to the $500 Intel Pentium 4 chip.

The use of DNA computing could finally dull Moore's Law. Leonard Aldeman, the father of the ground breaking DNA computing work has said that "DNA has been storing the blueprint of life for several billion years. Its powers are an untapped legacy for the 21st century." The relatively young topic is a different approach to technology then we're currently used to seeing. The new area of research bridges the study of enzymology, nanotechnology, synthetic chemistry, and computer science. Adleman believes, and hopes that the research he began over 15 years ago can unite the study of mathematics and biology and provide the kind of focused progress witnessed during the renaissance error by scientists like Leonardo DaVinci and Galileo.